476 ARTICLE Riparian forest structure and stream geomorphic condition: implications for flood resilience William S. Keeton, Erin M. Copeland, S. Mazeikaˇ P. Sullivan, and Mary C. Watzin Abstract: Managing riparian corridors for flood resilience requires understanding of linkages between vegetation condition and stream geomorphology. Stream assessment approaches increasingly use channel morphology as an indicator of stream condi- tion, with only cursory examination of riparian vegetation. Our research (i) examines relationships between stream geomorphic condition, as assessed by Rapid Geomorphic Assessment (RGA) scores, and riparian forest structure, and (ii) investigates scale dependencies in the linkages between land cover and stream geomorphology. We sampled vegetation structure and composition and assessed geomorphic condition at 32 stream reaches within the Lake Champlain Basin, USA. RGA scores were modeled as a function of structural attributes using classification and regression trees. Landsat coverages were used to delineate land uses within five nested spatial scales. Generalized linear models (GLM) evaluated relationships between land cover and RGA scores. Standard deviation of basal area partitioned the greatest variability in RGA scores, but dead tree density and basal area (positively) and shrub density (negatively) were also significant predictors. RGA was related to forest and agricultural cover at the two finest scales. Riparian forest structure is highly dynamic in relation to stand development and disturbance history; simple forest cover information does not capture these differences or their influences on stream geomorphic condition. Key words: stream ecosystems, riparian forest structure, stream geomorphology, watershed management, flood resilience. Résumé : L’aménagement de corridors riverains pour favoriser la résilience aux inondations exige qu’on comprenne les liens entre l’état de la végétation et la géomorphologie des ruisseaux. Les approches visant a` évaluer les cours d’eau font de plus en plus appel a` la morphologie du canal en tant qu’indicateur de l’état d’un cours d’eau et se limitent a` un examen superficiel de la végétation riveraine. Notre recherche examine (i) les relations entre la géomorphologie d’un cours d’eau, selon les résultats de la technique d’évaluation géomorphologique rapide (EGR), et la structure de la forêt riveraine et (ii) si les liens entre la couverture terrestre et la géomorphologie d’un cours d’eau dépend de l’échelle. Nous avons échantillonné la composition et la structure de la végétation et évalué l’état géomorphologique de 32 tronçons des cours d’eau dans le bassin du lac Champlain, aux États-Unis. Les résultats de l’EGR ont été modélisés en fonction des attributs structuraux a` l’aide d’arbres de régression et de classification. La couverture du satellite Landsat a servi a` délimiter l’utilisation des terres selon cinq échelles spatiales imbri- For personal use only. quées. Des modèles linéaires généralisés ont servi a` évaluer les relations entre la couverture terrestre et les résultats de l’EGR. L’écart-type de la surface terrière partageait la plus grande variabilité des résultats de l’EGR, mais la densité et la surface terrière des arbres morts (positivement) ainsi que la densité des arbustes (négativement) étaient aussi des prédicteurs significatifs. L’EGR était reliée au couvert forestier et agricole aux deux échelles les plus fines. La structure de la forêt riveraine est très dynamique relativement au développement et a` l’historique des perturbations du peuplement; l’information simple au sujet du couvert forestier ne détecte pas ces différences ou leurs influences sur la géomorphologie d’un cours d’eau. [Traduit par la Rédaction] Mots-clés : écosystème des cours d’eau, structure de la forêt riveraine, géomorphologie des cours d’eau, aménagement des bassins, résilience aux inondations. 1. Introduction et al. 2004; Sullivan and Watzin 2008; Sullivan 2012; Valero et al. 2015). While the structure of riparian forest stands is known to In the wake of two major weather events within the last five influence in-stream aquatic habitat characteristics and energy years (Tropical Storm Irene, 2011; Hurricane Sandy, 2012), manag- flows (Warren et al. 2016) through, for instance, the provisioning ing riparian corridors for flood resilience has become a pressing of shade and organic matter, most empirical studies on this topic concern in the northeastern United States (US Northeast) and in the US have been conducted in the western US (Bilby and Ward other regions. Developing a better understanding of linkages be- Can. J. For. Res. Downloaded from cdnsciencepub.com by UNIVERSITY OF VERMONT on 05/26/21 1991; Fetherston et al. 1995; Naiman et al. 1998, 1999; Warren et al. tween vegetation condition and stream channel geomorphology 2013). Recent research has increased our understanding of ripar- is vital for informing these efforts. Stream condition assessment ian forest structure and dynamics in stream systems in the US approaches (both physical and biological) in the US and globally Northeast (e.g., Hughes and Cass 1997; Keeton et al. 2007; Laser have used channel geomorphology as an indicator of ecological et al. 2009; Morris et al. 2010; Bechtold et al. 2016). Riparian forest condition (Rowntree and Wadeson 2000; Maryland Department of structure is likely to have important reciprocal relationships with Natural Resources (MDDNR) 2001; Southerland 2003; Sullivan various aspects of stream channel geomorphology and ecological Received 29 July 2016. Accepted 6 December 2016. W.S. Keeton and E.M. Copeland. Rubenstein School of Environment and Natural Resources, University of Vermont, Burlington, VT 05405, USA. S.M.P. Sullivan. School of Environment and Natural Resources, The Ohio State University, 2021 Coffey Road, Columbus, OH 43210, USA. M.C. Watzin. College of Natural Resources, Room 2028A, Biltmore Hall, North Carolina State University, Raleigh, NC 27695, USA. Corresponding author: William S. Keeton (email: [email protected]). Copyright remains with the author(s) or their institution(s). Permission for reuse (free in most cases) can be obtained from RightsLink. Can. J. For. Res. 47: 476–487 (2017) dx.doi.org/10.1139/cjfr-2016-0327 Published at www.nrcresearchpress.com/cjfr on 16 January 2017. Keeton et al. 477 condition, influencing the entrainment and distribution of large These include incising, widening, and eventually re-stabilizing in wood (LW) and rootwads (Gurnell et al. 2005; Warren et al. 2009), a new geometry, as conceptualized in channel evolution models bank erosion and sedimentation (Zaimes and Schultz 2015), chan- presented by Schumm (1977) and Schumm et al. (1984). nel geometry and adjustment (e.g., channel widening, changes in The RGA is a semiquantitative evaluation designed to synthe- meanders) (Hession et al. 2003; Sullivan et al. 2004; Sweeney et al. size channel condition using field indicators related to channel 2004), chemical water quality (Souza et al. 2013; Fernandes et al. 2014), adjustment (e.g., bank erosion, sediment deposition, channel avulsion, and energy dynamics (Gregory et al. 1991; Naiman et al. 1999). entrenchment, connectivity to floodplain, etc.; Vermont Department Thus, riparian forest condition may represent an indirect (through of Environmental Conservation (VTDEC) 2003; Sullivan et al. 2006) linkages with geomorphology), though important, control on both and stage of channel evolution. RGA scores are useful for assessing in-stream habitat characteristics and stream-flow regimes (Thorne stream geomorphic condition at the reach scale (typically 101–102 m) 1990; Johnson et al. 1995; Shields and Gippel 1995; Tabacchi et al. and have been shown to be reflective of more quantitative chan- 2000; Rahmeyer et al. 1999). Riparian forest control on stream geo- nel evaluations and to help define past and (or) current mecha- morphology has previously been termed “forced morphology” nisms of channel adjustment. Higher RGA scores correlate with (Montgomery and Buffington 1997). more stable channel configurations (VTDEC 2003) and have been Understanding relationships between riparian forest structure shown to be tightly linked to aquatic biota (Sullivan et al. 2004, and stream condition is particularly important in human-dominated 2006; Sullivan and Watzin 2008). landscapes such as northern New England, where land-use history The RGA is one of several techniques used to prioritize stream has dramatically altered forest age-class distributions (Lorimer reaches for restoration, protection, and management in the US 2001; Lorimer and White 2003; Nislow 2005), as well as the distri- (MDDNR 2001; VTDEC 2003; Rosgen 1996) and abroad (e.g., Cohen bution and structure of forest vegetation (Foster et al. 1998; et al. 1996; Rowntree and Wadeson 2000; King and Day 2002). Other Cogbill 2000; Fuller et al. 2004). These changes are likely to have methods focus more directly on aquatic biota (e.g., Karr 1981; altered structure-related riparian forest functions such as organic Davis and Simon 1995; Barbour et al. 1999) rather than on geomor- matter input and storage (Keeton et al. 2007; Warren et al. 2009), phology. As a reach-scale assessment, the RGA is typically used in flood dispersion (Rahmeyer et al. 1999), sediment retention and conjunction with watershed-scale analyses. While limited by its transport (Fetherston et al. 1995), and the composition and abun- semiquantitative nature, the RGA is actively used by natural re- dance of riverine biotic communities (Jones et al.